Research on Cargo Category Detection by Improved YOLOv5m

doi:10.3778/j.issn.1002-8331.2206-0293

Abstract

Abstract: Aiming at the problem of cargo category detection when trucks are put into storage, this paper proposes a method that can be used to detect the carried cargo online. This method is based on YOLOv5m and DeepSort algorithms. In the backbone network of YOLOv5m, the standard convolution is replaced by a depthwise separable convolution, which reduces the amount of model parameters and improves the model inference speed. The activation function SiLU is replaced by GELU, and the idea of random regularity is introduced. Residual structure is used to further improve network performance. Loss function CIoU is replaced by EIoU to improve regression accuracy. AdamW optimizer is used to improve parameter update. Finally, the self-made dataset is used for training and experiments. The experimental results show that the improved YOLOv5m model has the characteristics of higher accuracy, less calculation and faster detection, which can better meet the detection of goods in the warehouse environment.

Key words: YOLOv5m, DeepSort, objection detection, GELU, AdamW, EIoU

摘要： 针对货车入库时对货物类别检测问题，提出一种可以用于在线检测携带货物的方法。该方法基于YOLOv5m和DeepSort算法，在YOLOv5m的主干网络中以深度可分离卷积替换标准卷积，降低模型参数量，提高模型推理速度；激活函数SiLU替换为GELU，引入随机正则的思想；融入倒置残差结构，进一步提高网络性能；损失函数CIoU替换为EIoU，提高回归精度；采用AdamW优化器改善参数更新。最后通过自制数据集进行训练和实验，实验结果表明，改进后的YOLOv5m模型具有精度高、计算量小和检测速度快的特点，能够更好地满足仓储环境下的货车入库货物检测。

关键词: YOLOv5m, DeepSort, 目标检测, GELU, AdamW, EIoU

SUN Yuan, LI Weixiang, ZHOU Haijun. Research on Cargo Category Detection by Improved YOLOv5m[J]. Computer Engineering and Applications, 2023, 59(16): 205-211.

孙圆, 李为相, 周海军. 用于货物类别检测的改进YOLOv5m方法研究[J]. 计算机工程与应用, 2023, 59(16): 205-211.

References

[1] 贾峰，田野.基于自动化立体仓库中货物自动识别技术分析[J].冶金管理，2020（19）：67-68.
JIA F，TIAN Y.Analysis based on automatic identification technology of goods in automated stereoscopic warehouse[J].Metallurgical Management，2020（19）：67-68.
[2] 杨欢，幸芦笙.自动化立体仓库中货物自动识别技术[J].江西科学，2019，37（2）：287-292.
YANG H，XING L S.Automatic identification technology of goods in automated warehouse[J].Jiangxi Science，2019，37（2）：287-292.
[3] REDMON J，DIVVALA S K，GIRSHICK R B，et al.You only look once：unified，real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[4] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：6517-6525.
[5] REDMON J，FARHADI A.YOLOv3：an incremental improvement[J].arXiv：1804.02767，2018.
[6] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：optimal speed and accuracy of object detection[J].arXiv：2004.10934，2020.
[7] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision，2016：21-37.
[8] GIRSHICK R B，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[9] REN S Q，HE K M，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（6）：1137-1149.
[10] HE K M，GKIOXARI G，DOLLAR P，et al.Mask R-CNN[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision，2017：2980-2988.
[11] HENRIQUES J F，CASEIRO R，MARTINS P，et al.High-speed tracking with kernelized correlation filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2015，37（3）：583-596.
[12] WOJKE N，BEWLEY A，PAULUS D.Simple online and realtime tracking with a deep association metric[C]//Proceedings of the 2017 IEEE International Conference on Image Processing，2017：3645-3649.
[13] 何宇，李丹.基于YOLOv5和DeepSort的实时地铁客流量检测[J].电子测试，2022，36（2）：46-48.
HE Y，LI D.Real-time subway ridership detection based on YOLOv5 and DeepSort[J].Electronic Testing，2022，36（2）：46-48.
[14] 王晨，袁庆霓，白欢，等.面向仓储货物的轻量化目标检测算法[J].激光与光电子学进展，2022，59（24）：74-80.
WANG C，YUAN Q N，BAI H，et al.Lightweight object detection algorithm for warehousing goods[J].Progress in Laser and Optoelectronics，2022，59（24）：74-80.
[15] 黄凯文，凌六一，王成军，等.基于改进YOLO和DeepSORT的实时多目标跟踪算法[J].电子测量技术，2022，45（6）：7-13.
HUANG K W，LING L Y，WANG C J，et al.Real-time multi-target tracking algorithm based on improved YOLO and DeepSORT[J].Electronic Measurement Technology，2022，45（6）：7-13.
[16] 丁正道，吴红兰，孙有朝.基于YOLOv5+DeepSORT检测数据的车头时距混合分布模型研究[J].测控技术，2023，42（7）：56-64.
DING Z D，WU H L，SUN Y C.Research on time headway mixture distribution model based on YOLOv5+DeepSORT detection data[J].Measurement and Control Technology，2023，42（7）：56-64.
[17] HENDRYCKS D，GIMPEL K.Bridging nonlinearities and stochastic regularizers with Gaussian error linear units[J].arXiv：1606.08415，2016.
[18] ZHANG Y F，REN W Q，ZHANG Z，et al.Focal and efficient IOU loss for accurate bounding box regression[J].arXiv：2101.08158，2021.
[19] LOSHCHILOV I，HUTTER F.Decoupled weight decay regularization[C]//Proceedings of the 7th International Conference on Learning Representations，2019.
[20] HOWARD A G，ZHU M，CHEN B，et al.MobileNets：efficient convolutional neural networks for mobile vision applications[J].arXiv：1704.04861，2017.
[21] BEWLEY A，GE Z Y，OTT L，et al.Simple online and realtime tracking[C]//Proceedings of the 2016 IEEE International Conference on Image Processing，2016：3464-3468.
[22] 赵珊，刘子路，郑爱玲，等.基于MobileNetV2和IFPN改进的SSD垃圾实时分类检测方法[J].计算机应用，2022，42（S1）：106-111.
ZHAO S，LIU Z L，ZHENG A L，et al.Real-time classification and detection method of garbage based on SSD improved with MobileNetV2 and IFPN[J].Computer Application，2022，42（S1）：106-111.